Nemesysco's Layered Voice Analysis (LVA™)
From career choices to personal relationships, from science to art, from infancy to maturity – Understanding human emotions and motivations has always been a key challenge in almost every aspect of our lives. Undoubtedly, the better understanding of human emotions one has, the better he can react and interact…
Nemesysco’s technology was originally invented in 1997 by Amir Liberman, and is being developed ever since in consultation with experts from different relevant fields (psychology, criminology, phonetics, psychiatry, mathematics and others).
The technology is based on a proprietary set of vocal parameters found through research to correlate with key human emotions, and in various combinations to be able to identify deceptive intentions in “real life” scenarios. These vocal parameters were identified from a bank of audio files taken in different languages and a numerous settings, including police interrogations, call centers and controlled experiments.
Many of the parameters Nemesysco’s technology uses are new to the world of phonetics, and focus on the uncontrolled and yet phonetically unexplained properties of the human voice.
Layered Voice Analysis (LVA™) is Nemesysco's core technology adapted to meet the needs of various professional users, ranging from formal police investigations to security clearances, from vetting and recruitment needs to personality tests, and from call center’s quality monitoring to intelligence gathering and lawful interception multi-channel analysis.
In its security uses, LVA technology enables better understanding of your subject’s mental state and emotional reaction at a given moment by detecting the hidden emotional cues in his or her speech. The technology identifies various types of stress levels, cognitive processes, and emotional reactions that are reflected in different subtle properties of the voice. This information provides insights into the way your subject thinks, what troubles him, what excites him, what portions of his speech he is uncertain about, what questions require more of his attention, and what areas appear to be sensitive issues for the speaker.
LVA uses a unique mathematical process to detect different types of patterns and anomalies in the speech flow and classify them in terms of stress, excitement, confusion and other relevant emotional states, found through our 19 years of research to be highly correlated with these emotions.
Nemesysco’s Layered Voice Analysis is a completely new and proprietary technology and has no relations to the any old voice analysis technologies, nor does it uses any previously known method for detecting "voice stress". LVA does not perform "micro-tremors" (8-14 Hz) analysis in the voice, and can be used over telephone lines and with standard commercially available microphones.
The Human Speaking Mechanism
The human speaking process is one of the most complicated procedures the human body performs due to the number of muscles and physical apparatus involved, and the ways in which they need to be synchronized.
When a person speaks, air is pushed upward from the lungs into the vocal cords. Consequently, the vocal cords vibrate at a specific frequency and produce sound. The air continues to flow up toward the mouth where it is manipulated by the tongue, teeth and lips to produce sound streams that we recognize as words or phrases. The brain plays a leading role by orchestrating the entire process and monitoring what is said to ensure that the sound you emit is what your brain intended, is at the right volume to be heard by the intended listener, and is further accepted by the listener in the manner you intended it to be received.
LVA properties ignore the content of what your subject is saying and the language being used, and instead focus on the patterns and anomalies in the speech flow.
The “Emotional Diamond”
Nemesysco’s Emotional Diamond concept was developed in 2012 to simplify the visualization and understanding of the human emotional reaction complexity as detected by the LVA. It is built of 8 wings: “Energy” at the top, “Stress” at the bottom, “Logic and confidence” to the left, “Emotion” to the right, and 4 additional indications displaying levels of “Passion”, “Uneasy”, “Mental Effort” and “Concentration” (these are only the main 8 of the 14 emotional states LVA can detect).
The core principle behind the Emotional Diamond is that there are no good or bad emotions, as they are all context dependent. For example, we would expect to see stressful reactions when discussing intimidating events, and we will look for energetic and emotional reactions when we discuss prospects and achievements. When examining the emotional reactions over time, certain traits can be easily observed, and teach us about the true personality of the evaluated party.
LVA Emotion Detection Capabilities
The latest version of Layered Voice Analysis core technology (LVA7 CORE) released in 2014, gives more focus than previous versions to its unique abilities to explore and understand different emotions and personality traits. LVA’s emotion detection capabilities are in use today by commercial entities, such as call centers service assurance, HR recruiting, marketing research and even in medical assessment protocols for various needs. The use of LVA in these fields opens for the first time the gate for a wealth of information and the ability to scale and grade different emotions and track them down to the moment and reason of occurrence.
LVA and "Lie Detection"
Professionals in the field of lie detection know that there is no such thing as a “real” lie detector, since lying is not a unified reaction that is consistent at all times. Even the polygraph, the most known “Lie detector”, can only function properly in its pre-defined settings and in the relevant context. Simply connecting a polygraph to a person watching a video or performing any physical activity will render of course meaningless results.
Lying is a result of a deep logical process that is executed with a particular and significant intention. One might lie to protect himself from harm, while another might lie to gain profit, or even just to make a joke. Due to this variety in motivations and intentions, there is no fixed set of characteristics (physiological or psychological) that differentiate "lies" from "truth". However, LVA can detect and sharp change in the associated emotional reactions (a “jump out of homeostasis”) and assist you, the professional user, to expose the risk and uncover the truth.
Security Vs. Commercial uses
Intentions and emotional settings change according to the situation, and it is only expected that one will feel far less comfortable (or more “stressed”) when investigated by an officer of the law than when speaking with his friends. In such unique emotional settings, various emotions and other internal processes come to play.
LVA further provides you with tools to explore any identified piece of information in depth. LVA technology can be utilized both in real-time (for general overview of any subject/case/suspect/witness), and in offline mode (for more in-depth exploration), using recorded data from almost any source.
LVA based tools can be configured for different expected work settings, and has a built in mechanism to set its sensitivity automatically to meet different needs.
LVA Analysis Process
LVA7 has 3 basic formulas comprised of unique signal-processing algorithms that extract more than 150 parameters from each voice segment. These are further classified into 14 major categories of basic emotions. Depending on the goal of the analysis, several “Final Analysis” formulas can also be applied to the emotional parameter data: “Emotional Diamond Profile”, “Lie Stress” (“Risk Level” analysis), "Arousal Level", "Attention Level", “Anger or satisfaction detection“, “Personality”, and additional methods for veracity assessments.
This data can be further processed by statistical learning algorithms to predict the probability of deceptive or fraudulent intention in a sentence or for any other psychological analysis.
While the basic technology used in LVA is protected by 2 different patents, most of its internal processes are Nemesysco’s secret know-how.
LVA™ and Layered Voice Analysis™ are Nemesysco's trademarks.